Atlas

Benchmarks

← All benchmarks

Arena Vision — Diagram (No Style Control)

Multimodal

Raw Vision Arena rating for diagram-understanding prompts without style-control adjustment. Higher ratings indicate stronger preference in blind pairwise user votes.

Top models (higher is better)

ModelScore
Claude Fable 51357
Claude Opus 51348
Gemini 3.8 Flash1345
Claude Fable 5.11343
Opus 4.61340
Opus 4.71336
Muse Spark 1.31330
Gemini 3.7 Flash1329
Claude Sonnet 5.51329
GPT-5.41327
GPT-5.51325
Gemini 3.5 Flash1323
Qwen3.8-Max1323
GLM-5.3 Flash1317
Muse Spark 1.21317
Opus 4.81315
Gemini 3.6 Flash1314
Gemini 3.1 Pro Preview1311
Muse Spark1309
Step 5 Preview1308
Sonnet 4.61306
GPT-6.1 Sol1306
GPT-5.6 Sol1305
Muse Spark 1.11305
Grok 4.51304
Gemini 3 Pro Preview1301
Kimi K2.61298
GPT-6 Astra1297
DeepSeek V4.1 Flash1297
GPT-5.6 Terra1295
Sonnet 51293
Qwen3.7-Plus (source unspecified)1292
Gemini 3 Flash Preview1290
GLM 5V Turbo1286
Grok 4.61285
Qwen3.8 27B1285
Gemini 3.5 Flash-Lite1284
Gemma 4 31B IT1283
Kimi K2.51282
MiMo-V2.6-Pro1281
Loading Atlas data…